Recent understanding of neural stem cell biology and advances in the utilization of viral vectors and gene transfer techniques allow the application of these strategies to the repair of spinal cord injury. Multipotent neural stem cells are a promising source of transplants because they can be maintained and propagated in vitro, have the potential to differentiate into cell types characteristic of the host region, and can be genetically modified to express therapeutic genes that promote survival and regeneration. We have isolated and characterized spinal cord stem cells, have shown that their differentiation can be potentiated by selective extrinsic signals, and have demonstrated that when they are grafted into spinal cord they survive, and in the presence of neurotrophic factors, differentiate into neurons and glial cells. We propose in Aim 1 to transplant these cells into a unilateral hemisection site in adult spinal cord and determine their in vivo properties in terms of survival, differentiation and migration. We will examine the ability of the grafts to rescue injured neurons using stereological and image analysis methods and the effects of the transplant on regeneration of supraspinal neurons that are important for locomotion using retrograde and anterograde tracing methods. These experiments will establish the relative therapeutic properties of the grafted cells and the correlation between anatomical evidence of regeneration and behavioral evidence of function recovery. In Aim 2 we will compare the effectiveness of transplants modified to produce neurotrophic factors (NT3, BDNF) and an adhesion protein (LI). We will determine which of these grafts is the most effective in rescuing axotomized neurons in the Red Nucleus and promoting regeneration of rubrospinal, vestibulospinal, reticulospinal and corticospinal neurons. Recovery of motor and sensory function will be evaluated in one Project and correlated to the presence of the graft by re-lesion experiments. Physiological assessment of sensorimotor and Red Nucleus properties will be carried out in one Project using multi-electrode arrays. In Aim 3 we will test the ability of modified cells and delivery of therapeutic genes into transected adult spinal cord to promote regeneration of descending pathways sufficient to improve recovery of hindlimb function. This model will provide the most unequivocal and compelling evidence for the effectiveness of our strategy and its possible application in clinical treatment for spinal cord injury. We will use the protocols that produce the maximum rescue and regeneration and relate the anatomical and immunocytochemical analyses to recovery shown by specific behavioral tests. Physiological and biomechanical tests will be carried out in one Project to assess the extent of active projections and, if achieved, characterize weight support